论文标题
部分可观测时空混沌系统的无模型预测
Nature and Scalings of Density Fluctuations of Compressible MHD Turbulence with Applications to the Solar Wind
论文作者
论文摘要
储层计算是预测湍流的有力工具,其简单的架构具有处理大型系统的计算效率。然而,其实现通常需要完整的状态向量测量和系统非线性知识。我们使用非线性投影函数将系统测量扩展到高维空间,然后将其输入到储层中以获得预测。我们展示了这种储层计算网络在时空混沌系统上的应用,该系统模拟了湍流的若干特征。我们表明,使用径向基函数作为非线性投影器,即使只有部分观测并且不知道控制方程,也能稳健地捕捉复杂的系统非线性。最后,我们表明,当测量稀疏、不完整且带有噪声,甚至控制方程变得不准确时,我们的网络仍然可以产生相当准确的预测,从而为实际湍流系统的无模型预测铺平了道路。
The solar wind is a magnetized and turbulent plasma. Its turbulence is often dominated by Alfvénic fluctuations and often deemed as nearly incompressible far away from the Sun, as shown by in-situ measurements near 1AU. However, for solar wind closer to the Sun, the plasma $β$ decreases (often lower than unity) while the turbulent Mach number $M_t$ increases (can approach unity, e.g., transonic fluctuations). These conditions could produce significantly more compressible effects, characterized by enhanced density fluctuations, as seen by several space missions. In this paper, a series of 3D MHD simulations of turbulence are carried out to understand the properties of compressible turbulence, particularly the generation of density fluctuations. We find that, over a broad range of parameter space in plasma $β$, cross helicity and polytropic index, the turbulent density fluctuations scale linearly as a function of $M_t$, with the scaling coefficients showing weak dependence on parameters. Furthermore, through detailed spatio-temporal analysis, we show that the density fluctuations are dominated by low-frequency nonlinear structures, rather than compressible MHD eigen-waves. These results could be important for understanding how compressible turbulence contributes to solar wind heating near the Sun.